+// Computes the augmented script set for a code point, according to UTS 39 section 5.1.
+void SpoofImpl::getAugmentedScriptSet(UChar32 codePoint, ScriptSet& result, UErrorCode& status) {
+ result.resetAll();
+ result.setScriptExtensions(codePoint, status);
+ if (U_FAILURE(status)) { return; }
+
+ // Section 5.1 step 1
+ if (result.test(USCRIPT_HAN, status)) {
+ result.set(USCRIPT_HAN_WITH_BOPOMOFO, status);
+ result.set(USCRIPT_JAPANESE, status);
+ result.set(USCRIPT_KOREAN, status);
+ }
+ if (result.test(USCRIPT_HIRAGANA, status)) {
+ result.set(USCRIPT_JAPANESE, status);
+ }
+ if (result.test(USCRIPT_KATAKANA, status)) {
+ result.set(USCRIPT_JAPANESE, status);
+ }
+ if (result.test(USCRIPT_HANGUL, status)) {
+ result.set(USCRIPT_KOREAN, status);
+ }
+ if (result.test(USCRIPT_BOPOMOFO, status)) {
+ result.set(USCRIPT_HAN_WITH_BOPOMOFO, status);
+ }
+
+ // Section 5.1 step 2
+ if (result.test(USCRIPT_COMMON, status) || result.test(USCRIPT_INHERITED, status)) {
+ result.setAll();
+ }
+}
+
+// Computes the resolved script set for a string, according to UTS 39 section 5.1.
+void SpoofImpl::getResolvedScriptSet(const UnicodeString& input, ScriptSet& result, UErrorCode& status) const {
+ getResolvedScriptSetWithout(input, USCRIPT_CODE_LIMIT, result, status);
+}
+
+// Computes the resolved script set for a string, omitting characters having the specified script.
+// If USCRIPT_CODE_LIMIT is passed as the second argument, all characters are included.
+void SpoofImpl::getResolvedScriptSetWithout(const UnicodeString& input, UScriptCode script, ScriptSet& result, UErrorCode& status) const {
+ result.setAll();
+
+ ScriptSet temp;
+ UChar32 codePoint;
+ for (int32_t i = 0; i < input.length(); i += U16_LENGTH(codePoint)) {
+ codePoint = input.char32At(i);
+
+ // Compute the augmented script set for the character
+ getAugmentedScriptSet(codePoint, temp, status);
+ if (U_FAILURE(status)) { return; }
+
+ // Intersect the augmented script set with the resolved script set, but only if the character doesn't
+ // have the script specified in the function call
+ if (script == USCRIPT_CODE_LIMIT || !temp.test(script, status)) {
+ result.intersect(temp);
+ }
+ }
+}
+
+// Computes the set of numerics for a string, according to UTS 39 section 5.3.
+void SpoofImpl::getNumerics(const UnicodeString& input, UnicodeSet& result, UErrorCode& /*status*/) const {
+ result.clear();
+
+ UChar32 codePoint;
+ for (int32_t i = 0; i < input.length(); i += U16_LENGTH(codePoint)) {
+ codePoint = input.char32At(i);
+
+ // Store a representative character for each kind of decimal digit
+ if (u_charType(codePoint) == U_DECIMAL_DIGIT_NUMBER) {
+ // Store the zero character as a representative for comparison.
+ // Unicode guarantees it is codePoint - value
+ result.add(codePoint - (UChar32)u_getNumericValue(codePoint));
+ }
+ }
+}
+
+// Computes the restriction level of a string, according to UTS 39 section 5.2.
+URestrictionLevel SpoofImpl::getRestrictionLevel(const UnicodeString& input, UErrorCode& status) const {
+ // Section 5.2 step 1:
+ if (!fAllowedCharsSet->containsAll(input)) {
+ return USPOOF_UNRESTRICTIVE;
+ }
+
+ // Section 5.2 step 2
+ // Java use a static UnicodeSet for this test. In C++, avoid the static variable
+ // and just do a simple for loop.
+ UBool allASCII = TRUE;
+ for (int32_t i=0, length=input.length(); i<length; i++) {
+ if (input.charAt(i) > 0x7f) {
+ allASCII = FALSE;
+ break;
+ }
+ }
+ if (allASCII) {
+ return USPOOF_ASCII;
+ }
+
+ // Section 5.2 steps 3:
+ ScriptSet resolvedScriptSet;
+ getResolvedScriptSet(input, resolvedScriptSet, status);
+ if (U_FAILURE(status)) { return USPOOF_UNRESTRICTIVE; }
+
+ // Section 5.2 step 4:
+ if (!resolvedScriptSet.isEmpty()) {
+ return USPOOF_SINGLE_SCRIPT_RESTRICTIVE;
+ }
+
+ // Section 5.2 step 5:
+ ScriptSet resolvedNoLatn;
+ getResolvedScriptSetWithout(input, USCRIPT_LATIN, resolvedNoLatn, status);
+ if (U_FAILURE(status)) { return USPOOF_UNRESTRICTIVE; }
+
+ // Section 5.2 step 6:
+ if (resolvedNoLatn.test(USCRIPT_HAN_WITH_BOPOMOFO, status)
+ || resolvedNoLatn.test(USCRIPT_JAPANESE, status)
+ || resolvedNoLatn.test(USCRIPT_KOREAN, status)) {
+ return USPOOF_HIGHLY_RESTRICTIVE;
+ }
+
+ // Section 5.2 step 7:
+ if (!resolvedNoLatn.isEmpty()
+ && !resolvedNoLatn.test(USCRIPT_CYRILLIC, status)
+ && !resolvedNoLatn.test(USCRIPT_GREEK, status)
+ && !resolvedNoLatn.test(USCRIPT_CHEROKEE, status)) {
+ return USPOOF_MODERATELY_RESTRICTIVE;
+ }
+
+ // Section 5.2 step 8:
+ return USPOOF_MINIMALLY_RESTRICTIVE;
+}
+
+int32_t SpoofImpl::findHiddenOverlay(const UnicodeString& input, UErrorCode&) const {
+ bool sawLeadCharacter = false;
+ for (int32_t i=0; i<input.length();) {
+ UChar32 cp = input.char32At(i);
+ if (sawLeadCharacter && cp == 0x0307) {
+ return i;
+ }
+ uint8_t combiningClass = u_getCombiningClass(cp);
+ // Skip over characters except for those with combining class 0 (non-combining characters) or with
+ // combining class 230 (same class as U+0307)
+ U_ASSERT(u_getCombiningClass(0x0307) == 230);
+ if (combiningClass == 0 || combiningClass == 230) {
+ sawLeadCharacter = isIllegalCombiningDotLeadCharacter(cp);
+ }
+ i += U16_LENGTH(cp);
+ }
+ return -1;
+}
+
+static inline bool isIllegalCombiningDotLeadCharacterNoLookup(UChar32 cp) {
+ return cp == u'i' || cp == u'j' || cp == u'ı' || cp == u'ȷ' || cp == u'l' ||
+ u_hasBinaryProperty(cp, UCHAR_SOFT_DOTTED);
+}
+
+bool SpoofImpl::isIllegalCombiningDotLeadCharacter(UChar32 cp) const {
+ if (isIllegalCombiningDotLeadCharacterNoLookup(cp)) {
+ return true;
+ }
+ UnicodeString skelStr;
+ fSpoofData->confusableLookup(cp, skelStr);
+ UChar32 finalCp = skelStr.char32At(skelStr.moveIndex32(skelStr.length(), -1));
+ if (finalCp != cp && isIllegalCombiningDotLeadCharacterNoLookup(finalCp)) {
+ return true;
+ }
+ return false;
+}
+
+
projects / apple / icu.git / blobdiff